Method and apparatus for tissue region identification

CROSS-REFERENCE TO RELATED APPLICATIONS/INCORPORATION BY REFERENCE

p-0002This application makes reference to: U.S. patent application Ser. No. 13/671,190 filed on Nov. 7, 2012, which issued as U.S. Pat. No. 8,824,758 B2 on Sep. 2, 2014.

p-0003Each of the above referenced applications is hereby incorporated herein by reference in its entirety.

FIELD

p-0004Certain embodiments of the disclosure relate to pathology imaging. More specifically, certain embodiments of the disclosure relate to a method and apparatus for tissue region identification.

BACKGROUND

p-0005In the area of biology and medicine, understanding cells and their supporting structures in tissues and tracking their structure and distribution changes is crucial to advancing medical knowledge in disease diagnoses. Histology is the study of the microscopic anatomy of tissues and is essential in diagnosing disease, developing medicine and many other fields. In histology, thin slices of tissue samples are placed on one or more slides and then examined under a light microscope or electron microscope. Often, however, the tissue samples are placed in any portion of the slide and the slide itself may be oriented in different ways due to variable layouts in digital pathology systems, causing difficulty in accurate location of a target area on a slide. In addition, comparing an input image of a tissue sample to other input images of similar tissue samples becomes difficult because the tissue samples may be stretched or compressed and the resulting slide image becomes distorted.

p-0006Therefore there is a need in the art for a method and apparatus for tissue region identification in digital pathology.

p-0007Further limitations and disadvantages of conventional and traditional approaches will become apparent to one of skill in the art, through comparison of such systems with some aspects of the present disclosure as set forth in the remainder of the present application with reference to the drawings.

SUMMARY

p-0008An apparatus and/or method is provided for tissue region identification in digital pathology substantially as shown in and/or described in connection with at least one of the figures, as set forth more completely in the claims.

p-0009These and other features and advantages of the present disclosure may be appreciated from a review of the following detailed description of the present disclosure, along with the accompanying figures in which like reference numerals refer to like parts throughout.

BRIEF DESCRIPTION OF THE DRAWINGS

p-0010<figref idrefs="DRAWINGS">FIG. 1</figref> is a block diagram illustrating a tissue region identification module in accordance with an embodiment of the disclosure;

p-0011<figref idrefs="DRAWINGS">FIG. 2</figref> is a block diagram of a computer system for implementing a tissue region orientation module in accordance with embodiments of the present invention;

p-0012<figref idrefs="DRAWINGS">FIG. 3</figref> is a flow diagram illustrating a method for estimating a tissue region in a slide image according to exemplary embodiments of the present invention;

p-0013<figref idrefs="DRAWINGS">FIG. 4</figref> is a flow diagram illustrating a method for determining if a segment is background data according to exemplary embodiments of the present invention; and

p-0014<figref idrefs="DRAWINGS">FIG. 5</figref> is a flow diagram illustrating a method for separating multiple objects in an input tissue region according to exemplary embodiments of the present invention.

DETAILED DESCRIPTION

p-0015Certain implementations may be found in an apparatus and/or method for tissue region identification in digital pathology. According to one embodiment, a digital pathology image is segmented into a plurality of regions each containing structures or objects which may or may not be relevant tissue or background artifacts and the like. Each region is subjected to filtering where objects in the region are determined. If the objects are determined to be curvilinear segments, these objects are considered to be background data and not contain relevant data to the tissue sample. Once all curvilinear portions of the data are filtered out, artifacts and background are determined and removed from the remaining regions. Separate regions of interest are identified in the remaining regions by separating multiple objects in the tissue sample, thereby identifying a tissue region in a digital pathology image. As used herein, “curvilinear” means any image portion that is substantially one-dimensional, that is, a curved or straight line.

p-0016<figref idrefs="DRAWINGS">FIG. 1</figref> is a block diagram illustrating a tissue region identification module <b>100</b> in accordance with an embodiment of the disclosure. The tissue region identification module <b>100</b> comprises a segmentation module <b>102</b>, a background module <b>103</b>, a segment filter module <b>104</b>, and a region refinement module <b>106</b>. The tissue region identification module <b>100</b> has an input slide image <b>101</b> as input. According to some embodiments, the input slide image <b>101</b> is a digital pathology image which is stained with one of various color dyes to enhance cellular visibility.

p-0017The input slide image <b>101</b> is first divided into a plurality of segments by the segmentation module <b>102</b>. According to one embodiment, the segmentation module <b>102</b> divides the image <b>102</b> into X segments. Disparate regions are separated based on spatial proximity, as well as the textural and contrast characteristics in each of the X segments. As a result, tissue regions in the slide image <b>101</b> are considered foreground and segmented from regions of surrounding texture which are not connected to the tissue regions. In some embodiments, the input image <b>101</b> is 100,000 pixels by 100,000 pixels.

p-0018The background module <b>103</b> estimates background distribution from regions in the segmented image that are not considered to be foreground.

p-0019The segment filter module <b>104</b> determines which portions of the input slide image <b>101</b> are tissue regions and which are merely artifacts or weak tissue regions and are one-dimensional. The segment filter module <b>104</b> performs filtering on each region produced by the segmentation module <b>102</b>. The portions in each region which are deemed to be artifacts or weak tissue areas are filtered from the image. According to one exemplary embodiment, if a portion of the region contains curvilinear segments, then that portion is deemed to not contain a target area because, portions that are curvilinear are often artifacts.

p-0020Therefore, segment filter module <b>104</b> filters out from the input slide image <b>101</b> any portions which resemble curvilinear line segments. A curvilinear line segment is any segment that resembles a one dimensional curve or line. According to one embodiment of the present invention, a portion of the input slide image can be determined to be curvilinear by applying an area ratio test.

p-0021The area ratio test is applied by finding the minimum x and y and the maximum x and y of a portion of the segment input to the segment filter module <b>104</b>, when viewing the segment on the x-y plane. The min/max x, y provide the coordinates of a bounding box for enclosing the suspected curvilinear portion of the current segment. An area is then computed for the bounding box region. An area is further computed for the full length of the input region

p-0022The ratio of the bounding box area to the area of the input region is compared to a threshold value. If the ratio is smaller than the threshold value, the segment within the bounding box is determined to be either a line or the structure has a very sparse point distribution with little useful information, and is deemed an artifact. Therefore it is discarded. However, if the ratio is equal to or higher than the threshold value, the bounded segment is significant and is kept for further processing and is not deemed an artifact.

p-0023The segment filter module <b>104</b> further can fit a particular portion or segment to a curve to determine if the portion is curvilinear. If the segment fits a curve or a straight line very closely, the segment is discarded as lacking important information.

p-0024The region refinement module <b>106</b> identifies the background and isolates the target data region. According to some embodiments, the region refinement module <b>106</b> also performs multiple object separation where there is greater than one region of interest. According to some embodiments, the region refinement algorithm is described in related application.

p-0025According to one embodiment, the region refinement module <b>108</b> performs multiple object separation. The region refinement module <b>108</b> calculates pixel separation between boundaries of objects in the input image <b>101</b>. If the number of pixels between two objects is below a particular threshold value, the two objects are considered as a single object. If the number of pixels between the two objects is greater than the predetermined threshold value, the objects are separated.

p-0026However, if two objects in a region are separated by a diagonal distance, for example, a 45 degree separation, a more complex method is used to determine the pixel distance between the two objects. In one embodiment, a tight region is computed around each object, and this boundary is used as a separation marker. The distance from the boundary to another object is examined to determine whether the pixel distance is greater or smaller than the predetermined threshold value.

p-0027The region refinement module <b>108</b> refines the non-filtered portions of the plurality of regions into a target data region <b>110</b>. The target data region <b>110</b> is the region in the input image <b>101</b> where the tissue sample, or the significant portions of the tissue region, is identified.

p-0028<figref idrefs="DRAWINGS">FIG. 2</figref> is a block diagram of a computer system <b>200</b> for implementing the orientation module <b>100</b> in accordance with embodiments of the present invention. The computer system <b>200</b> includes a processor <b>202</b>, a memory <b>204</b> and various support circuits <b>206</b>. The processor <b>202</b> may include one or more microprocessors known in the art, and/or dedicated function processors such as field programmable gate arrays programmed to perform dedicated processing functions. The support circuits <b>206</b> for the processor <b>202</b> include microcontrollers, application specific integrated circuits (ASIC), cache, power supplies, clock circuits, data registers, input/output (I/O) interface <b>208</b>, and the like. The I/O interface <b>208</b> may be directly coupled to the memory <b>204</b> or coupled through the supporting circuits <b>206</b>. The I/O interface <b>208</b> may also be configured for communication with input devices and/or output devices <b>210</b>, such as, network devices, various storage devices, mouse, keyboard, displays, sensors and the like.

p-0029The memory <b>204</b> stores non-transient processor-executable instructions and/or data that may be executed by and/or used by the processor <b>202</b>. These processor-executable instructions may comprise firmware, software, and the like, or some combination thereof. Modules having processor-executable instructions that are stored in the memory <b>204</b> comprise the region identification module <b>220</b>, further comprising the segmentation module <b>222</b>, the segment filter module <b>226</b>, the background module <b>224</b> and the region refinement module <b>228</b>.

p-0030The computer <b>200</b> may be programmed with one or more operating systems (generally referred to as operating system (OS) <b>214</b>, which may include OS/2, Java Virtual Machine, Linux, Solaris, Unix, HPUX, AIX, Windows, Windows95, Windows98, Windows NT, and Windows 2000, Windows ME, Windows XP, Windows Server, among other known platforms. At least a portion of the operating system <b>214</b> may be disposed in the memory <b>204</b>. In an exemplary embodiment, the memory <b>204</b> may include one or more of the following: random access memory, read only memory, magneto-resistive read/write memory, optical read/write memory, cache memory, magnetic read/write memory, and the like, as well as signal-bearing media, not including non-transitory signals such as carrier waves and the like.

p-0031<figref idrefs="DRAWINGS">FIG. 3</figref> is a flow diagram illustrating a method <b>300</b> for estimating a tissue region in a slide image according to exemplary embodiments of the present invention. The method <b>300</b> is an implementation of the region identification module <b>220</b> as executed by the CPU <b>202</b> of the computer system <b>200</b>, shown in <figref idrefs="DRAWINGS">FIG. 2</figref>.

p-0032The method <b>300</b> starts at step <b>302</b> and proceeds to step <b>304</b>. At step <b>304</b>, the segmentation module <b>222</b> segments an input slide image into a plurality of regions based on detected structures and objects in the image.

p-0033The method then proceeds to step <b>306</b>, where structures which are curvilinear (that is, substantially one-dimensional) are determined to be part of the background and are filtered out from the foreground by the segment filter module <b>226</b>. As described above with reference to <figref idrefs="DRAWINGS">FIG. 1</figref>, the segment filter module <b>226</b> performs various different tests such as the Area Ratio test, fitting to curves, and the like to determine whether a structure or segment is substantially one-dimensional and therefore contains only background data, or contains relevant tissue sample data.

p-0034At step <b>308</b>, the background module <b>224</b> detects and removes artifacts across the plurality of regions. At step <b>310</b>, the remaining area is isolated as the target area, because all background information has been identified and removed. Thus, what is left is considered foreground data and relevant. Thus, the sample tissue region has been estimated and identified. The method then ends at step <b>312</b>.

p-0035<figref idrefs="DRAWINGS">FIG. 4</figref> is a flow diagram illustrating a method <b>400</b> for determining if a segment is background data according to exemplary embodiments of the present invention. The method <b>400</b> is an implementation of the segment filter module <b>226</b> as executed by the CPU <b>202</b> of the computer system <b>200</b>, shown in <figref idrefs="DRAWINGS">FIG. 2</figref>.

p-0036The method <b>400</b> starts at step <b>402</b> and proceeds to step <b>404</b>. At step <b>404</b>, a bounding box is determined for a structure or region that is possibly curvilinear. The bounding box is determined by determining the minimum x, minimum y and maximum x and maximum y of the structure in the region.

p-0037At step <b>406</b>, the area of the bounding box is computed. At step <b>408</b>, the area of the segment within which the structure lies is computed. The method proceeds to step <b>410</b> where a ratio of the area of the bounding box to the area of the segment is calculated. This ratio will give a measure of an estimate of how much area the structure takes up of the region.

p-0038At step <b>410</b>, if the ratio is very small, i.e., smaller than a particular threshold value, the method moves to step <b>412</b>. At step <b>412</b>, the segment filter module <b>226</b> determines the structure is curvilinear, and is background data and is filtered out of the final target area. If the ratio is not less than a threshold value, the structure is deemed as not curvilinear, and is considered relevant and the segment is retained as part of the final area. The method <b>400</b> ends at step <b>416</b>.

p-0039<figref idrefs="DRAWINGS">FIG. 5</figref> is a flow diagram illustrating a method <b>500</b> for separating multiple objects in an input region according to exemplary embodiments of the present invention. The method <b>500</b> is an implementation of a portion of the region refinement module <b>228</b> as executed by the CPU <b>202</b> of the computer system <b>200</b>, shown in <figref idrefs="DRAWINGS">FIG. 2</figref>.

p-0040The method <b>500</b> starts at step <b>502</b> and proceeds to step <b>504</b>. At step <b>504</b>, the region refinement module <b>228</b> computes the distance between a first and second object in the input region. When the two objects are vertically or horizontally displaced, the distance between the two objects is determined as the pixel distance between the boundaries of each object in a horizontal or vertical direction. However, if the objects are displaced at an angle from other, tight regions are computed around each object and distances are calculated from the borders of these regions.

p-0041At step <b>506</b>, the region refinement module <b>228</b> determines whether the distance is smaller than a threshold value. If the distance is smaller than a threshold value, the first and second objects are considered as single objects at step <b>508</b>. If the distance is not smaller than a threshold value, the first and second objects are treated as separate objects and distinct regions of interest at step <b>510</b>. The method ends at step <b>512</b>.

p-0042Accordingly, the present disclosure may be realized in hardware, or a combination of hardware and software. The present disclosure may be realized in a centralized fashion in at least one computer system or in a distributed fashion where different elements may be spread across several interconnected computer systems. Any kind of computer system or other apparatus adapted for carrying out the methods described herein may be suited. A combination of hardware and software may be a general-purpose computer system with a computer program that, when being loaded and executed, may control the computer system such that it carries out the methods described herein. The present disclosure may be realized in hardware that comprises a portion of an integrated circuit that also performs other functions.

p-0043The present disclosure may also be embedded in a computer program product, which comprises all the features enabling the implementation of the methods described herein, and which when loaded in a computer system is able to carry out these methods. Computer program in the present context means any expression, in any language, code or notation, of a set of instructions intended to cause a system having an information processing capability to perform a particular function either directly or after either or both of the following: a) conversion to another language, code or notation; b) reproduction in a different material form.

p-0044While the present disclosure has been described with reference to certain embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the scope of the present disclosure. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the present disclosure without departing from its scope. Therefore, it is intended that the present disclosure not be limited to the particular embodiment disclosed, but that the present disclosure will include all embodiments falling within the scope of the appended claims.